speaker_01.json_confirmed
>
Agent EngineeringLondon ConferenceThe first ever conference covering every engineering behind AI agents.
>London isEngineering
PromptContextHarnessMemoryLoopGraphInferenceCode
Deep technical talks for CTOs, agent builders, AI researchers, and technical leaders engineering Agentic AI.
Founders Edition · Limited seats
Everyman, Canary Wharf, London · Friday 16 October 2026Add to calendar
until doors open
Agent Engineering HQSan Francisco·London
EARLY ACCESS OPEN
Early access list and speaker applications open now
>
Speakers
The lineup is announced on a rolling basis. Here is who is confirmed so far.
>lineup5 confirmed · limited spots
speaker_06
$ agenteng community --london
Organized by Agent Engineering HQ with London Agentic AI, the UK's biggest Agentic AI community of 5,000+ engineers.
$ agenteng audience --london
The London community behind the conference: engineers, founders and researchers shipping agent systems.
5,000+
community members
>cat README.md
What is Agent Engineering Conference?
// our definition
Agent Engineering is the discipline of designing, building, evaluating, and operating AI agents reliably in production. It focuses on the engineering aspects of the Agent Development Lifecycle, including prompt, context, harness, inference, memory, evaluation, and code engineering practices.
AgentEng (Agent Engineering Conference) is the first conference dedicated to the engineering practices behind production-grade AI agents of every kind, from coding agents and research agents to customer-facing agents, domain-specific agents, and autonomous multi-agent systems.
As agents become core to how software and businesses are built, the engineer's role shifts from writing code to directing, reviewing, and orchestrating the agents that do the work.
Agents have their own development cycle, the Agent Development Lifecycle (ADLC), covering build, test, deploy, and monitor with iteration and governance running throughout. The conference covers every aspect of the ADLC and the tools, frameworks, and infrastructure that support it.
We bring together practitioners doing this in production, working with real constraints, real trade-offs, and real systems.
$ agenteng --list-disciplines
Code
Context
Evaluation
Graph
Memory
>
From Prompt Engineering to Agent Engineering
AI system engineering keeps specializing in production. In practice, these disciplines converge in one place: Agent Engineering.
discipline::01
Prompt Engineering
discipline::02
Context Engineering
discipline::03
Harness Engineering
discipline::04
Eval Engineering
discipline::05
Memory Engineering
discipline::06
Inference Engineering
discipline::07
Loop Engineering
discipline::08
Agentic Engineering
discipline::09
Code Engineering
discipline::10
Protocol Engineering
discipline::11
Graph Engineering
discipline::12
Search Engineering
discipline::∞
*.Engineering
convergence::active
Agent Engineering
Prompts, context, harnesses, evals, memory, inference, loops, protocols, graphs, search, and code converge in production systems engineering.
$ all_roads --lead-to agent_engineering
Prompt Engineering
Context Engineering
Harness Engineering
Eval Engineering
Memory Engineering
Inference Engineering
Loop Engineering
Agentic Engineering
Code Engineering
Protocol Engineering
Graph Engineering
Search Engineering
discipline::∞
... *.Engineering
convergence::result
Agent Engineering
Production agentic systems require more than prompts alone. They depend on prompts, context, harnesses, evals, memory, inference, loops, protocols, graphs, search, and code working together as one engineering discipline.
>agenteng cfp --open
Speak at AgentEng London
We are looking for AI engineers, AI researchers, CTOs, and CEOs who ship agents in production and know the engineering practices that make them reliable. We want highly technical talks with live demos, real architecture, and lessons from running agentic systems at scale.
A curated programme with a limited number of technical talks. Speaker applications are open.
who we are looking for
- AI engineers, AI researchers, CTOs, and CEOs who ship agents in production
- Engineers who know the hardcore practices that make agentic systems reliable
- Practitioners who can speak to architecture, components, and production trade-offs
>
what we are looking for
- Highly technical talks with live demos and working components
- Real architecture from production systems, including proprietary applications
- Lessons learned from running agentic systems at scale
- Practical insights that benefit a senior engineering audience
>
talk formats
Technical Talk25 minutes · one engineering discipline per talk
Audience Q&A10 minutes · scheduled into every slot
Curated ProgrammeA limited number of talks, chosen by the programme team
We would love to hear about the systems you build and operate.
topics we want on stage
Pick a theme that fits your work
These are the engineering themes we are programming. If your system or your company operates in one of these areas, please get in touch.
>
Prompt Engineering
Writing, structuring, and optimizing prompts and system instructions that hold up in production.
fit: Teams working on prompt optimization, structured outputs, and prompt evaluation tooling.
>
Context Engineering
Building, compressing, and routing the right context to the model.
fit: Companies working on context optimisation and context windowing.
>
Harness Engineering
The infra around the model: tool runtimes, sandboxes, agent loops.
fit: Companies innovating on agent harnesses and coding harnesses.
>
Memory Engineering
Long-term, short-term, episodic, and semantic memory in agents, including tool search and retrieval over that memory.
fit: Agent-memory and memory-architecture companies, vector databases, and retrieval platforms.
>
Inference Engineering
Serving, scaling, and optimising model inference for agentic workloads.
fit: Inference platforms, model-serving companies, and teams optimising agent latency and cost.
>
Eval Engineering
Measuring agent behaviour: regressions, traces, replay, scoring.
fit: Companies building agent observability and evaluation platforms.
>
Loop Engineering
Designing the agent run loop: control flow, stopping conditions, retries, self-correction, and long-running autonomous loops.
fit: Teams building agent runtimes, loop controllers, and self-correcting autonomous systems.
>
Agentic Engineering
Automating the entire software development lifecycle and engineering with AI agents.
fit: Teams and companies automating software engineering itself with AI agents.
>
Code Engineering
Improving the coding capabilities of models and agents: code generation, coding agents, and agentic coding workflows.
fit: Companies building coding agents, code-generation models, and AI developer tools.
>
Protocol Engineering
Standing up the agent protocol layer: MCP, ACP, A2A, A2UI, and the interfaces agents talk over.
fit: Teams building protocol servers and clients, and companies working on agent interoperability.
>
Graph Engineering
Building agents with graphs: graph databases, knowledge graphs, and graph-shaped state and execution.
fit: Graph database vendors, knowledge-graph platforms, and teams modelling agent workflows as graphs.
>
Search Engineering
Agentic search techniques, and improving agent behaviour through better retrieval and ranking.
fit: Search and retrieval platforms, and teams tuning search to make their agents measurably better.
Each slot is 25 minutes of talk followed by 10 minutes of questions from the audience. We ask speakers to take those questions on stage and to leave a way of being contacted on their final slide. Attendees pay to hear from you and to ask you things, and the second part matters as much as the first.
No Sessionize or CFP portal. A short form and a few sentences is all we need. We read every application, and will be in touch if yours is selected.
Speaker applications and the early access list are open now.
>
Agent Engineering London · Friday 16 October 2026
Join a curated room of technical builders at Everyman, Canary Wharf, London for a focused day on engineering agentic AI systems.
1
Day
200
Attendee Cap
12
Engineering Themes
The Venue
Everyman Canary Wharf, a boutique cinema at Crossrail Place, London E14 5AR.
The Room
Talks are presented on a cinema screen with 7.1 audio, with premier armchair and sofa seating throughout the auditorium.
Getting There
Canary Wharf station on the Jubilee and Elizabeth lines, with DLR at Canary Wharf, Poplar, and West India Quay. Lift access to the cinema level, with accessible spaces in the auditorium.
Everyman, Canary Wharf → London → 16 October 2026
>Next stop: San Francisco
>
Who This Conference Is For
This is a technical conference for people who ship agent systems. Expect architecture, debugging, evaluations, infra trade-offs, and production reliability conversations.
@builder
Agent Engineers & Builders
Hands-on engineers designing runtimes, tool chains, eval harnesses, and on-call operations for agents.
@ml-engineer
AI / ML Engineers
Shipping model integration, inference pipelines, eval workflows, and guardrails in production systems.
@platform
Developer Tool & Platform Builders
Building frameworks, SDKs, orchestration layers, observability, and deployment infrastructure for agent systems.
@product
Product Engineers & Technical PMs
Owning API-level product surfaces, eval targets, and agent behavior quality across releases.
@founder
Founders & Technical Leaders
Leading technical teams building agent-native products or internal automation platforms.
@researcher
Researchers & Applied Scientists
Driving applied research in multi-agent architectures, reliability, and evaluation methodology.
>
Why AgentEng Is Different
--engineering-first
Engineering-First
--discipline-led
Discipline-Led Agenda
--separate-tracks
Programme and Sponsorship, Separately
--attendee-led
Attendee-Led
>
Program Themes
A curated technical program, built around deep system walkthroughs, practical lessons, and hard engineering trade-offs.
--optimize
Agent Optimization
Automatically optimize agents across all layers: prompts, RAG, protocols, memory, and context.
--design-ax
Agent Experience (AX)
Designing machine-readable interfaces, environments, and feedback loops that shape Agent Experience.
--future-code
Agentic Coding
How software development evolves as agents become first-class contributors to the SDLC.
--business-models
Agentic Business Models
Defining the business models for agentic AI. Does SaaS still work, or is FDE the only way forward? Explore and share business models.
>cat agent-engineering/README.md
Agent Engineering is an emerging discipline.
A manifesto, a mindset, and a set of engineering themes for practitioners building serious agentic systems.
// organisers
Organised by Superagentic AI
Built on community, shaped with practitioners, speakers, and collaborators.
Superagentic AI
Advancing the Agent Engineering discipline and practice
leadership@agenteng ~
$agenteng whoami --chair
[ Shashi Jagtap ]
Conference Chair & Curator · Ex-Apple Engineer (6 years) · Organizer, London Agentic AI
Building the bridge from Developer Experience to Agent Experience. Transatlantic connections between London and Silicon Valley, with programming shaped by a broader technical community.
>
Frequently Asked Questions
Everything you need to know about the Agent Engineering Conference
>agenteng agenda --preview
A preview of the day
Tentative
A limited number of technical talks, each covering a different agent engineering discipline. Every talk is 25 minutes followed by 10 minutes of questions from the audience. Times and themes will change as the programme is confirmed.
Limited
technical talks
25 min
per talk
10 min
Q&A after every talk
1
day, 09:00 to 16:30
Every talk includes questions
Ten minutes of every slot belongs to the audience. The time is built into the schedule rather than borrowed from it, so it does not disappear when the day runs late. Where a room has no questions, the host puts them to the speaker.
Every slot on the schedule is a technical talk followed by audience questions. We have kept the day to that one format.
How questions are run
- One question per person. Ask again once everyone else has had a turn.
- Ninety seconds to ask it. The host will step in and move things on at ninety seconds.
- No follow-up debates. If an answer opens a longer discussion, continue it with the speaker afterwards.
- Questions are taken in order. Submit through Slido during the talk, or raise a hand if the queue is empty.
$ cat agenda.txt
16 October 2026
- 08:00Doors open, registration and coffeeOpen
- 09:00Welcome and openingOpen
- 09:10Talk 1theme: Context Engineering25 min + 10 min Q&A
- 09:50Talk 2theme: Harness Engineering25 min + 10 min Q&A
- 10:30Morning breakBreak
- 10:50Talk 3theme: Memory Engineering25 min + 10 min Q&A
- 11:30Talk 4theme: Eval Engineering25 min + 10 min Q&A
- 12:10Talk 5theme: Inference Engineering25 min + 10 min Q&A
- 12:50LunchLunch
- 13:50Talk 6theme: Protocol Engineering25 min + 10 min Q&A
- 14:30Talk 7theme: Code Engineering25 min + 10 min Q&A
- 15:10Afternoon teaBreak
- 15:30Talk 8theme: Agentic Engineering25 min + 10 min Q&A
- 16:10Closing remarksClose
- 16:30Networking and drinkstheme: Same venue and nearby pubsNetworking
Draft agenda: sessions, themes, and times will change as the programme is confirmed.
>
Let's Connect
Suggest a technical speaker, enquire about sponsorship, or get in touch with the team.
Early bird registration open · speaker applications open · sponsorships open
$ Quick response guaranteed